arena ยท diff
git:20261028.933dd9d to git:20261101.8be107b
16 added, 30 removed. Audit A to A.
---
name: arena
- description: "Compare competing design or implementation sketches when a consequential choice cannot be resolved cheaply from existing evidence."
+ description: "Build and compare competing design or implementation sketches when a consequential choice cannot be resolved cheaply from existing evidence. Use architect for a single design."
---
# Arena
Follow the consuming project's instructions and requested scope; use the
[portable defaults](../orbyn-workflow/references/project-policy.md) only where guidance is missing.
- State the artifact and a few observable criteria before generating alternatives.
- Use at most two candidates initially, each with its own output path and rationale.
- The owning lead uses named native Orbyn Architect/Engineer assignments under the
- [workflow](../orbyn-workflow/SKILL.md), with disjoint candidate paths. Keep at most
- two active children across the task; compare sequentially when slots are occupied.
- A child forbidden delegation compares sketches directly within its owned scope.
- Follow [coordination](../orbyn-workflow/references/project-policy.md#delegation-and-host-configuration) for user-requested project
- sessions; candidate comparison alone does not authorize new chats. Never launch
- hidden CLI model sessions or recursive tournaments.
+ ## Method
- Give candidates the same task and constraints without steering them to a preferred
- answer. The lead reads every returned artifact and its rationale. Judge against requirements and maintainability, not
- model reputation or agreement count. Retain one coherent design and incorporate
- only compatible improvements. Do not average incompatible assumptions.
+ 1. **State the decision** in one sentence, and the artifact. Freeze the inputs, hard constraints and a few observable scoring criteria before generating alternatives.
+ 2. **Build at most two candidates initially,** preferably differing on the uncertain design choice. Give each its own output path and rationale; shared generated files and databases need their own owner. Give both the same task and constraints without steering them toward a preferred answer.
+ 3. **Compare in a table:** constraint, candidate A evidence, candidate B evidence and remaining uncertainty. A hard constraint disqualifies a candidate before aesthetic or convenience scoring. Judge against requirements and maintainability, not model reputation or agreement count.
+ 4. **Inspect every returned artifact** and its rationale yourself rather than trusting its author's summary.
+ 5. **Select one coherent contract,** or report that neither meets the criteria. Incorporate only compatible improvements; never average incompatible assumptions or merge incompatible lifecycles. Preserve useful alternatives and rejected rationale.
+ 6. **Stop when the decision is supported.** The lead reviews the synthesized diff and performs authorized proportional verification. Return the selected artifact, why it won, rejected ideas, remaining uncertainty and the smallest unresolved check to the architecture or implementation owner.
- Record the selected approach, useful alternatives, rejected ideas and remaining
- uncertainty. The lead reviews the synthesized diff and performs authorized proportional
- verification;
- candidate agreement is not correctness evidence. Stop when the decision is made.
+ For a live-update decision, feed both sketches the same reconnect, ordering and resource constraints and the same disconnect, duplicate and out-of-order trace. Compare a concrete failure trace and maintenance consequence alongside the happy-path demo; settle delivery correctness from the trace, not a screenshot.
- ## Compare the same user outcome
+ ## Delegation limits
- For a live-update decision, give both sketches the same reconnect, ordering and resource constraints. Compare a concrete failure trace and maintenance consequence, not only the happy-path demo. Return one selected contract to architecture; this skill does not own an ongoing implementation tournament.
+ The owning lead uses named native Orbyn Architect/Engineer assignments under the [workflow](../orbyn-workflow/SKILL.md), with disjoint candidate paths. Keep at most two active children across the task; compare sequentially when slots are occupied. A child forbidden delegation compares sketches directly within its owned scope.
- ## Operational method
+ If the host cannot supply permitted delegates, compare sequential sketches directly and disclose that the review was not independent. Unavailable independent/model-family review remains a limit, not permission to change defaults or run a hidden panel.
- Use the [local operating method](references/operating-method.md) for the execution
- sequence, concrete failure case, expected result and stopping condition. It is
- Orbyn-authored guidance; historical source notices remain in provenance.
+ Follow [coordination](../orbyn-workflow/references/project-policy.md#delegation-and-host-configuration) for user-requested project sessions; candidate comparison alone does not authorize new chats. Never launch hidden CLI model sessions or recursive tournaments. Hand off after the decision, since an ongoing implementation tournament is out of scope. Report two agents agreeing as agreement, never as verified correctness.
- When comparing scored candidates, keep task acceptance visible but judge-only
- ranking details, model identities and condition metadata outside candidate scope.
- Use neutral labels and audit leakage under the
- [evaluation method](../orbyn-skill-creator/references/evaluations.md#candidate-context-and-judge-leakage).
- Unavailable independent/model-family review remains a limit, not permission to
- change defaults or run a hidden panel.
+ ## Scored candidates
+
+ When comparing scored candidates, keep task acceptance visible but keep judge-only ranking details, model identities and condition metadata outside candidate scope. Use neutral labels and audit leakage under the [evaluation method](../orbyn-skill-creator/references/evaluations.md#candidate-context-and-judge-leakage).